Chemical synthesis and biological activities of 16α-derivatives of 5α-androstane-3α,17β-diol as antiandrogens

Article abstract of DOI:10.1016/j.bmc.2007.02.007

In our efforts to develop compounds with therapeutic potential as antiandrogens, we synthesized a series of 5α-androstane-3α,17β-diol derivatives with a fixed side-chain length of 3-methylenes at C-16α, but bearing a diversity of functional groups at the

Full text of DOI:10.1016/j.bmc.2007.02.007

Bioorganic & Medicinal Chemistry 15 (2007) 3003–3018
Chemical synthesis and biological activities of 16a-derivatives of
5a-androstane-3a,17b-diol as antiandrogens
*
´
Jenny Roy, Rock Breton, Celine Martel, Fernand Labrie and Donald Poirier
Medicinal Chemistry Division, Oncology and Molecular Endocrinology Research Center,
Centre Hospitalier Universitaire de Que´bec (CHUQ), Pavillon CHUL and Universite´ Laval, Que., Canada G1V 4G2
Received 12 September 2006; revised 2 February 2007; accepted 8 February 2007
Available online 13 February 2007
Abstract—In our efforts to develop compounds with therapeutic potential as antiandrogens, we synthesized a series of 5a-andro-
stane-3a,17b-diol derivatives with a fixed side-chain length of 3-methylenes at C-16a, but bearing a diversity of functional groups
at the end. Among these, the chloride induced the best antiproliferative activity on androgen-sensitive Shionogi cells. Substituting
the OH at C-3 by a methoxy group showed the importance of the OH. Moreover, its transformation into a ketone increased the
androgen receptor (AR) binding but decreased the antiproliferative activity and induced a proliferative effect on Shionogi cells.
These results confirm the importance of keeping a 5a-androstane-3a,17b-diol nucleus instead of a dihydrotestosterone nucleus. Var-
iable side-chain lengths of 2-, 3-, 4-, and 6-methylenes at C-16a were investigated and the optimal length was found to be 3-meth-
ylenes. Although exhibiting a weak AR binding affinity, 16a-(30-chloropropyl)-5a-androstane-3a,17b-diol (15) provided an
antiproliferative activity on Shionogi cells similar to that of pure non-steroidal antiandrogen hydroxy-flutamide (77% and 67%,
respectively, at 0.1 lM). The new steroidal compound, 15, thus constitutes a good starting point for development of future antian-
drogens with a therapeutic potential against prostate cancer.
ꢀ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
the interaction of T and DHT with the AR. Androgen
ablation therapy has been shown to produce the most
beneficial responses in multiple settings in prostate
cancer patients.8 Several studies have reported that a
combination therapy of orchidectomy with an antian-
drogen, to inhibit the action of adrenal androgens,
significantly prolongs the survival.7–13 Since prostate
cancer is so highly sensitive to androgens, the antiandro-
gen used must be a compound having high specificity
and affinity for the AR while not possessing any andro-
genic, estrogenic, progestational, glucocorticoid or any
other hormonal and antihormonal activities.14,15
Prostate cancer is the most frequently diagnosed cancer
in men in the United States, accounting for 33% of all
cancers. It is estimated that 232,090 new cases of pros-
tate cancer will be diagnosed and 30,350 men will die
from this disease in 2005.1 Androgens testosterone (T)
and dihydrotestosterone (DHT) play an important role
in the development, growth, and progression of prostate
cancer.2–5 Androgen receptor (AR) binds the male sex
steroids, DHT and T, and regulates genes for male dif-
ferentiation and development.6 Therefore, mutations in
the AR gene may lead to several diseases or conditions
like prostate cancer or the androgen insensitivity syn-
drome.7 Since an essential step in the action of andro-
gens in target cells is binding to the receptor, a logical
approach for neutralizing the action of androgens is
the use of antiandrogens or compounds which prevent
Since even the best treatment of advanced or metastatic
prostate cancer can only prolong life with minimal or no
possibility of cure,9,16–20 it is important to increase the
efficiency of known treatments. One of the strategies
investigated is to develop a much more potent antian-
drogen than flutamide, a compound known to have a
weak AR binding affinity.14 An interesting improvement
was the development of selective androgen receptor
modulators (SARMs), which required the synthesis of
numerous non-steroidal compounds.21,22 Several steroid
scaffolds, such as cyproterone acetate,23 spironolac-
tone24 and 4-aza-heterocycle steroid,25 have also been
Keywords: Prostate cancer; Steroid; Androstane; Androgen receptor;
Chemical synthesis.
*
Corresponding author. Tel.: +1 418 654 2296; fax: +1 418 654
2761; e-mail: Donald.Poirier@crchul.ulaval.ca
0968-0896/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.02.007

3004
J. Roy et al. / Bioorg. Med. Chem. 15 (2007) 3003–3018
reported as antiandrogens in the past, but their develop-
ment was less impressive than non-steroidal
antiandrogens.
2. Results and discussion
2.1. Chemical synthesis
Following a study of the inhibition of the steroidogenic
enzyme type 3 17b-hydroxysteroid dehydrogenase,26 we
recently reported an interesting antiandrogenic profile
for the synthetic 16a-androstane derivative 1 (16a-(30-
bromopropyl)-5a-androstane-3a, 17b-diol) (Fig. 1).
Furthermore, its antiproliferative activity on androgen-
sensitive (AR+) Shionogi cells seems to be mediated
through direct interaction with AR as supported by its
AR binding affinity. Interestingly, it does not show sig-
nificant affinities for estrogen, glucocorticoid, and pro-
gestin receptors.26 With the aim of extending this
exploratory work, a series of 16a-peptidosteroids repre-
sented by structure 2 was synthesized using solid-phase
synthesis in parallel fashion.27 The screening of the gen-
erated model libraries revealed interesting preliminary
structure–activity relationship (SAR) related to their
antiproliferative activities on Shionogi cells, but these
activities were weak and not mediated by AR. We then
decided to extend our SAR study by focusing more clo-
sely on the lead compound 1. In this article, we report
the synthesis of androstane derivatives 3, which are var-
iously substituted at the end of the C-16-a side-chain of
3-methylenes or differently modified at C-3 to modulate
the biological activities. Focusing on two optimal substi-
tutions, another series of 5a-androstane-3a,17b-diol
derivatives bearing a bromoalkyl or chloroalkyl chain
of variable length at C-16a (compounds 4) was synthe-
sized. In addition to the chemical synthesis, proliferative
and antiproliferative activities on Shionogi cells as well
as the binding affinities for AR were determined. A
molecular modeling study was also performed in an at-
tempt to analyze interactions between a representative
compound and AR.
Dihydrotestosterone (5) was the starting material for the
synthesis of all compounds. First, we synthesized steroid
precursor 9 through a sequence of reactions described in
Scheme 1. The carbonyl group of 5 was protected as ke-
tal and the 17b-hydroxy group of 6 was oxidized with
TPAP and NMO to obtain ketone 7. An alkylation in
alpha position of the carbonyl was then performed to
give 8, a mixture of 16a-allyl and 16b-allyl stereoisomers
1
(90/10, evaluated by the H NMR signal of CH3-18).
After purification by chromatography, a mixture of
the major isomer a and minor isomer b (2%) was stere-
oselectively reduced using LiAlH4 at low temperature to
afford the secondary alcohol 9. As previously observed
for related compounds,26,28,29 NMR signals at C-17 con-
firmed the C-16a and 17b orientation of the allyl and
OH groups, respectively.
For the first part of our study, precursor 9 was submit-
ted to an oxidative hydroboration yielding diol 10.
Subsequently, several derivatives having a different func-
tionality at the end of a propyl side-chain were gener-
ated using strategies reported in Scheme 2. Chloride 11
was obtained from 10 after ketal hydrolysis and substi-
tution of the hydroxy group using CCl4 and PPh3. Un-
der these conditions, the hindered secondary alcohol at
the C-17b position was not reactive. Diol 12 was easily
obtained after deprotection of 10. The primary alcohol
of 12 was either substituted by an iodide giving 13 or
reduced to give triol 14. We selected potassium tri-sec-
butyl-borohydride (K-Selectride) as the reducing agent,
because this reagent30 gave mainly the 3a-OH isomer,
1
which was confirmed using H and 13C NMR data re-
ported in the literature.31,32 Alcohol 14 was substituted
Figure 1. Chemical structures of 5a-androstane-3a,17b-diol derivatives 1 and 2 previously prepared as potential antiandrogens and general structures
of analogues 3 and 4 synthesized to extend our SAR study. The stereogenic centers are illustrated only for steroid 1, but they are the same for all
other steroid derivatives reported in this paper.

J. Roy et al. / Bioorg. Med. Chem. 15 (2007) 3003–3018
3005
Scheme 1. Reagents and conditions: (a) (CH2OH)2, p-TSA, benzene, Dean-Stark apparatus, reflux, 24 h; (b) TPAP, NMO, molecular sieves, CH2Cl2,
rt, 1 h; (c) i—LDA (DIPA, BuLi, THF, 0 ꢁC), ii—CH2@CHCH2Br, THF, ꢀ78 to 0 ꢁC, 24 h; (d) LiAlH4, THF, ꢀ78 ꢁC, 6 h.
Scheme 2. Reagents and conditions: (a) i—BH3ÆTHF, 0 ꢁC, 5 h; ii—H2O2, NaOAc, H2O, 0 ꢁC to rt, 3 h; (b) HCl, acetone, rt, 1 or 2 h; (c) PPh3, CCl4,
CH2Cl2, 0 ꢁC to rt, 1 or 2 h; (d) PPh3, I2, CH2Cl2, imidazole, rt, 2 h; (e) K-Selectride, THF, ꢀ78 ꢁC, 1 h; (f) TBAF, THF, reflux, 16 h; (g) PPh3, CBr4,
CH2Cl2, 0 ꢁC to rt, 2 h; (h) NaN3, CH3CN, 80 ꢁC, 24 h; (i) Pd/C, MeOH, H2, 3 h; (j) KSCN, EtOH, reflux, 16 h.

3006
J. Roy et al. / Bioorg. Med. Chem. 15 (2007) 3003–3018
by a chloride to afford 15, which was substituted by a
fluoride to obtain 16. Iodide 17 was generated directly
from alcohol 14, because it was not possible to reduce
13 into 17. Indeed, this kind of iodoalkyl is sensitive
to K-Selectride reducing conditions. The intermediate
compound 12 was also substituted using CBr4 and
PPh3 to produce bromide 18. This compound was
substituted by an azide to give 19, which was reduced
at the C-3 position giving 20. Catalytic hydrogenation
of azide group provided the free amine 21. The bromide
of 18 was also substituted by KSCN in ethanol to give
22 and this ketone reduced to generate alcohol 23.
ketal hydrolysis, and the substitution of the primary
alcohol using CCl4 and PPh3. At this point, part of ke-
tones 32 and 33 was reduced at C-3 to give alcohols 34
and 35. Compounds 38 and 39 bearing a bromoethyl
chain were synthesized from 36, which was available in
our laboratory.27 This precursor was acetylated and
the TBDMS protecting group was thereafter removed
in a THF solution of HF/pyridine leading to hydroxyl
derivative 37. This alcohol was oxidized using TPAP
and NMO, and the acetyl group was removed with
potassium carbonate to give 38. To obtain compound
39, the TBDMS of precursor 36 was hydrolyzed under
acid conditions.
We next tried introducing an aromatic group at the end of
the 16a side-chain by preparing compounds 25 and 26
(Scheme 3). The secondary alcohol of 18 was first pro-
tected as a tert-butyldimethylsilyl ether derivative 24
using TBDMS-OTf and 2,6-lutidine. The bromide of 24
was substituted by a 4-bromobenzyloxy group and the
TBDMS intermediate was hydrolyzed to afford 25,
which, after carbonyl reduction with K-Selectride, affor-
ded 26. The methylation of the OH at position 3 of the ste-
roid was also investigated. To obtain 28, ketone 24 was
reduced and the intermediate 27 was methylated using
NaH and MeI. This SN2 reaction also gives the iodo ana-
logue. The mixture of bromide/chloride and iodo deriva-
tives was then hydrolyzed to remove the TMDBS group,
giving two new target compounds 28 and 29.
2.2. Biological activity
The antiproliferative activity of synthesized compounds
on androgen-sensitive mammary carcinoma Shionogi
cells is first reported as the percentage (%) of inhibition
(Table 1). It corresponds to the ability of a compound to
inhibit the proliferation induced by 0.3 nM of the natu-
ral potent androgen DHT. The proliferative activity is
reported as the percentage (%) of cell stimulation in-
duced by the tested compounds relative to stimulation
(100%) induced by 0.3 nM DHT. Moreover, it is impor-
tant to determine whether or not these compounds bind
to androgen receptor (AR), in order to find out whether
the antiproliferative activity is mediated by this receptor.
For the second part of our study, we needed to prepare
compounds with various side-chain lengths at C-16a.
The synthesis of the compounds with the two longer
side-chains required a Grubbs’s metathesis between the
key intermediate 9 and alcohols protected or not as ben-
zyl esters (Scheme 4). Using a second-generation ruthe-
nium catalyst33–35 allowed us to generate in reasonable
yields compounds 30 and 31 with 6- and 4-carbon chain,
respectively. Chlorides 32 and 33 were generated by the
following sequence of reactions: a double bond hydroge-
nation, a subsequent ester hydrolysis (for 31 only), a
First of all, substituting the bromide on the side-chain
end with a diversity of functional groups leads to inter-
esting SAR information. Clearly the substitution of the
bromide atom of 1 by a more polar group such as OH,
NH2, and SCN (compounds 14, 21, and 23) is detrimen-
tal to the antiproliferative activity. The same result is
also observed with a very bulky group such as the bro-
mobenzyloxy of 26. Among the series of halogeno deriv-
atives 1 and 15–17, chloride 15 induced the higher
antiproliferative activity at 0.1 lM. Although halogens
are good leaving groups, the results obtained with 1
Scheme 3. Reagents and conditions: (a) i—2,6-lutidine, TBDMS-OTf, CH2Cl2, ꢀ78 ꢁC, 3 h; ii—HCl (10% aq); (b) AgOTf, 4-bromobenzyl alcohol,
2,6-di-tert-butyl-4-methyl-pyridine, CH2Cl2, 0 ꢁC to rt, 24 h; (c) HCl, MeOH, CH2Cl2, rt, 2 h; (d) K-Selectride, THF, ꢀ78 ꢁC, 1 or 2 h; (e) NaH, MeI,
15-crown-5, THF, 0 ꢁC to rt, 24 h.

J. Roy et al. / Bioorg. Med. Chem. 15 (2007) 3003–3018
3007
Scheme 4. Reagents and conditions: (a) Grubbs’s catalyst, CH2Cl2, reflux, 24 h; (b) H2, Pd/C, EtOAc, rt, 24 h; (c) NaOH (10% aq), MeOH, rt, 1.5 h;
(d) HCl, acetone, rt, 1 or 2 h; (e) PPh3, CCl4, CH2Cl2, 0 ꢁC to rt, 1 h; (f) K-Selectride, THF, ꢀ78 ꢁC, 1 h; (g) Ac2O, pyridine, DMAP, rt, 2 h; (h) HF/
pyridine, THF, rt, 60 h; (i) TPAP, NMO, molecular sieves, CH2Cl2, rt, 2 h; (j) K2CO3, MeOH, reflux, 2 h; (k) HCl, MeOH, rt, 2 h.
and 15–17 suggest that these compounds do not work by
formation of a covalent bond with AR or another tar-
get. Like chloride 15, azide 20 also showed a good anti-
proliferative activity at 0.1 lM. Indeed, the mean result
of three experiments with 15 (67 5%) is not signifi-
cantly different from that obtained with 20 (58 9%).
Among all these new 3a-OH-steroids, none produced
proliferative activity at 0.1 lM, except compound 26.
The binding affinity of all these 3a-OH derivatives for
AR was also evaluated but none showed an important
affinity (1–17% at 0.1 lM and 4–53% at 1 lM).
ketones 11, 13, 19, and 22 were weaker (21–36% at
0.1 lM), but are representative of compounds with both
agonist and antagonist activities. Indeed, this conclusion
is well-illustrated on Figure 2 where the antiproliferative
and proliferative activities of ketone 11 and hydroxy-flu-
tamide (OH-Flu),36–38 expressed as cell DNA content,
are reported at different concentrations. Thus, the pure
antiandrogen OH-Flu inhibits the proliferation of
Shionogi (AR+) cells induced by DHT (OH-
Flu + DHT) in a concentration-dependent fashion, up
to 100% at the higher concentrations, but does not stim-
ulate the cell proliferation when tested alone (OH-Flu)
at any concentration. On the other hand, ketone 11 only
partially inhibits the cell proliferation induced by DHT
(11 + DHT), down to ꢁ40% at the highest concentra-
tion, but clearly stimulates cell proliferation when tested
alone (up to ꢁ52% at 1 lM). As a result, the two curves
corresponding to antiproliferative (11 + DHT) and pro-
liferative (11 only) activities converge to a same level of
cell growth, indicating mixed activities. Thus, the pres-
ence of a carbonyl group at position 3 of the androstane
derivatives leads to a compound with mixed androgenic
and antiandrogenic activities, confirming the impor-
tance of keeping the 3a-OH.
In the light of these results, we decided to evaluate the
importance of the 3a-OH group of these new derivatives
by performing two modifications. First, substituting the
3a-alcohol by a methoxy group (28 and 29) induced a
drop of antiproliferative activity and binding affinity
for AR. Second, with the aim of increasing the antago-
nistic effect and AR binding affinity, we prepared and
tested a series of ketones. As expected, the presence of
a carbonyl at C3 greatly improved the AR binding.
These better affinities were however associated with a
proliferative activity at 0.1 lM ranging from 9% to
153% (100% for DHT at 0.3 nM). The increase of bind-
ing and proliferative activity was the least important for
alcohol 12 but remarkable for 3-keto-steroid 25. In fact,
the proliferative activity of 25 (153% at 0.1 lM) was
similar to that of 26 (163%), the 3a-OH analogue. Com-
pared to that of 25 or DHT, the proliferative effects of
After we determined that a 3a-OH and a Cl atom at the
end of the 16a side-chain formed an interesting combi-
nation, we wanted to find the optimal side-chain length
(2-, 3-, 4- or 6-methylenes), the one that induces the best

3008
J. Roy et al. / Bioorg. Med. Chem. 15 (2007) 3003–3018
Table 1. Antiproliferative activity, proliferative activity, and AR binding affinity
OH
(CH ) R
2 3
X
aPercentage of antiproliferative activity on Shionogi (AR+) cells at 0.1 and 1 lM of tested compounds. The cells were stimulated with 0.3 nM DHT.
SEM 6 2.7% for the first value, representing an experiment performed in triplicate. Data in parentheses are from two other experiments performed
3 years later in triplicate (SEM 6 3.3%). Data in [ ] represent means SEM of the three experiments.
bPercentage of proliferative activity on Shionogi (AR+) cells at 0.1 lM of tested compounds. The stimulation induced by 0.3 nM of androgen DHT
was set as 100%. SEM 6 3.5% of one experiment in triplicate.
cPercentage of binding affinity on human androgen receptor at 0.1 and 1 lM of tested compounds. The binding of R1881 (1 lM) was set as 100%.
SEM 6 2.2% of one experiment in triplicate.
dHydroxy-flutamide (pure antiandrogen).
eMethyltrienolone (synthetic potent androgen).
fTestosterone (natural androgen).
gDihydrotestosterone (natural androgen).
antiproliferative activity (Fig. 3). Compound 15 with a
side-chain length of 3-methylenes displayed the greater
potency. In fact, at concentrations of 0.1 and 1 lM, this
chloride derivative inhibited by 77% and 97%, respec-
tively, the DHT stimulation. The antiproliferative activ-
ity was however lower with both a shorter and a longer
side-chain. As reported in Figure 4, the pattern of re-
sponse of the 3a-OH-steroid 15 is comparable to the
one obtained with the non-steroidal pure antiandrogen
OH-Flu. Thus, 15 inhibits the proliferation of Shionogi
(AR+) cells induced by DHT (15 + DHT) in a concen-
tration-dependent fashion, up to 100% at the higher
concentrations, but does not stimulate the cell prolifera-
tion when tested alone (15 only). Furthermore, these
two compounds have a similar binding affinity for AR
(33% and 36% at 1 lM for OH-Flu and 15, respectively).
AR, but instead an effect not mediated by AR. We then
tested ketone 11, alcohol 15, OH-Flu, and DHT in the
non-androgen-sensitive (ARꢀ) PC-3 cells. These human
prostatic carcinoma cells do not respond to androgens
and glucocorticoids, and are useful in assessing their
response to chemotherapeutic agents.39 As shown in
Figure 5, the cell growth was not reduced by a treatment
with ketone 11, alcohol 15, androgen DHT or antian-
drogen OH-Flu at concentrations tested. Thus, 11 and
15 are non-cytotoxic compounds in PC-3 (ARꢀ) cells
suggesting that the antiproliferative effect observed in
Shionogi (AR+) cells by 15 is mediated by AR despite
its low binding affinity.
2.3. Molecular modeling
In order to understand the antiproliferative activity
demonstrated by 16a-(30-chloropropyl)-5a-androstane-
3a, 17b-diol (15), we tried to identify the preferred
binding position of lead compound 1 within the AR li-
gand-binding domain (LBD) by modeling and energy
Since the AR binding affinity of 3a-OH-steroids is low
and does not correlate well with the biological results,
it is possible that the antiproliferative activity is not
the result of an antiandrogenic activity, mediated by

J. Roy et al. / Bioorg. Med. Chem. 15 (2007) 3003–3018
3009
Figure 2. Antiproliferative and proliferative effects induced by the
3-keto-steroid 11 on Shionogi (AR+) cells. The concentration of DHT
is 0.3 nM. The pure antiandrogen OH-Flu was used as a reference
compound. Data are expressed as means SEM of one experiment in
triplicate.
Figure 4. Antiproliferative and proliferative effects induced by the 3a-
OH-steroid 15 on Shionogi (AR+) cells. The concentration of DHT is
0.3 nM. The pure antiandrogen OH-Flu was used as a reference
compound. Data are expressed as means SEM of one experiment in
triplicate.
Figure 5. Proliferative effect induced by 3-keto-steroid 11, 3a-OH-
steroid 15, OH-Flu, and DHT on PC-3 (ARꢀ) cells. Data are expressed
as means SEM of one experiment in triplicate. Significantly different
from control (CTL): *(p < 0.01); **(p < 0.05).
Figure 3. Antiproliferative activity on Shionogi (AR+) cells as a
function of the side-chain length at C-16a. The cells were simulta-
neously incubated with 0.3 nM DHT. SEM 6 4% of one experiment in
triplicate. *Alcohol 15 is significantly different from all other com-
pounds at 0.1 lM (p < 0.01).
it. When binding an agonist, the top of the LBP is closed
by an a-helix structure at the carboxyl-terminal end of
the receptor, helix 12 (H12).7,41–43 A precise positioning
of this helix is required for activation of AR41 and, as
shown by crystallographic studies made on the human
estrogen receptor, a different positioning of H12 could
account for the functional differences between agonists
and antagonists.44 This position shift would prevent
coactivators from binding to the receptor. Structural
modeling of 1 bound inside the LBP of AR and compar-
ison with crystal structures of AR-LBD in complex with
minimization (Fig. 6). The AR-LBD is mainly com-
posed of a-helices arranged as a three-layered antiparal-
lel a-helical sandwich, a fold common to all steroid
receptors.40 The ligand-binding pocket (LBP) is mainly
composed of hydrophobic residues, the side-chains of
which can easily adopt variable positions in order to bet-
ter fit the hydrophobic core of the steroid and stabilize

3010
J. Roy et al. / Bioorg. Med. Chem. 15 (2007) 3003–3018
group has a lone pair of electrons, it could act as a
hydrogen bond acceptor able to establish a strong inter-
action with a charged amino acid, Arg752, like DHT
does.43 Thus, the specific interaction involving the C-3
ketone may orient the steroid core similarly to DHT
and thus reduce the efficiency of the alkyl chain in repo-
sitioning H12.
3. Conclusions
We have synthesized a novel series of antiandrogens by
adding a short side-chain at position 16a of a 5a-andro-
stane-3a,17b-diol nucleus. Among the series of func-
tional groups attached at the end of the alkyl side-
chain, we found that an azide or halogen best improves
the antiproliferative activity on androgen-sensitive
(AR+) Shionogi cells, especially when this element of
diversity is a chloride. Moreover, the optimal side-chain
length was found to be 3-methylenes. However, all of
these 3a-OH-steroids display a weak binding affinity
for the AR. We also showed the importance of a 3a-hy-
droxy group since its replacement by a methoxy de-
creases biological activity. Moreover, the presence of a
ketone at position 3 brought a better binding to AR,
while unfortunately inducing proliferation of Shionogi
(AR+) cells. These opposite effects certainly explain the
poor antiandrogenic activity of these 3-keto-steroids at
higher concentrations. Clearly, the 3a-OH-steroids and
the 3-keto-steroids work differently. The latter have a
good AR binding affinity, but work as mixed antago-
nist/agonist compounds, whereas the former have a
low AR binding affinity and a profile of activity similar
to that of the non-steroidal antiandrogen OH-Flu.
Among the compounds tested, the 3a-OH-steroid 15
with a chloropropyl 16a-side-chain displays the highest
antiandrogenic activity on Shionogi cells. Its inhibition
of DHT-induced cell stimulation is comparable to that
obtained with the non-steroidal antiandrogen OH-Flu.
Furthermore, compound 15 shows no proliferative
agonistic activity on Shionogi (AR+) cells and does
not reduce the proliferation of PC-3 (ARꢀ) cells.
Figure 6. Hypothetical position of 3a-OH-steroid 1 in the human
androgen receptor ligand binding site. Schematic views (b-face of the
steroid nucleus (a) and about 90ꢁ rotated view (b)) of 1 in the optimal
binding position. Putative hydrogen bonds established by the ligand
with the receptor (residues Arg752, Asn705, and Thr877) are indicated
by broken green lines. Possible contacts between the C16a-side-chain
of the ligand and the Phe876 residue of helix 11 (see text) are indicated
by broken pink lines. Carbon atoms are depicted in white, oxygen
atoms in red, nitrogen atoms in blue, and bromine atom in orange. The
figures were generated with Swiss-PdbViewer.46
agonist ligands, DHT and testosterone,43 suggest that 1
is well stabilized through numerous van der Waals con-
tacts with hydrophobic residues delineating the LBP and
by hydrogen bonds established at both extremities of its
steroidal nucleus. Its 16a-(30-bromopropyl) side-chain is
oriented toward the main chain of helix 11 and clashes
with the side-chain of one of its residues, Phe876. How-
ever, So¨derholm et al.45 reported the presence of a small
favored volume for acceptor interactions close to the
solvent accessible surface, located near the amide group
of residue Leu880 and the carbonyl oxygen of Phe876.
The possibility thus exists that this interference with he-
lix 11 has an impact on the positioning of the C-terminal
end of the receptor and causes a misplacement of H12
leading to the generation of the antagonist-like confor-
mation of AR. These interactions could explain the
antagonistic effect of alcohols 1 and 15. Moreover, the
agonistic effect of the latter’s 3-keto analogue, com-
pound 11, could be explained by a higher AR binding
affinity due to the presence of a ketone function at C-
3, like in DHT. Since the oxygen atom of this ketone
Based on modeling studies we tried to explain how com-
pound 1 antagonizes the AR. It was proposed that the
bromide of the 16a chain of the steroid likely makes
contacts with residues of helix 11, leading to a different
positioning of helix 12 and thus preventing coactivators
to bind to the receptor. Additional studies will however
be necessary to better understand the mechanism of ac-
tion of compounds 1 and 15. Moreover, a better under-
standing of the key interactions between 15 and AR
could contribute to the design of a second generation
of this new family of antiandrogens with improved
biological activity.
4. Experimental
4.1. Chemical synthesis
4.1.1. General. Dihydrotestosterone (DHT) was ob-
tained from Steraloids (Newport, RI, USA) and

J. Roy et al. / Bioorg. Med. Chem. 15 (2007) 3003–3018
3011
chemical reagents were purchased from Sigma–Aldrich
Canada Ltd. (Oakville, ON, Canada). The usual sol-
vents were obtained from Fisher Scientific (Nepean,
4.1.3. Synthesis of 3-dioxolane-5a-androstan-17-one (7).
To a solution of alcohol 6 (5.2 g, 15.6 mmol) in dry
CH2Cl2 (70mL) under argon were added molecular
˚
´
ON, Canada) and VWR International (Montreal,
sieves (4 A) (7.8 g) and 4-methylmorpholine-N-oxide
Qc, Canada) and were used as received. Anhydrous
solvents were purchased from Aldrich and VWR in
SureSeal bottles, which were conserved under positive
argon pressure. Tetrahydrofuran (THF) was distilled
from sodium/benzophenone ketyl under argon. Flash
(2.7 g, 23.3 mmol) and the mixture was stirred for
15 min at rt. Tetrapropylammonium perruthenate
(273 mg, 0.8 mmol) was added and the solution was
stirred for 1 h. The resulting mixture was filtered on sil-
ica gel column, using hexanes/acetone (80:20) as eluent
to give 5.1 g (98%) of ketone 7. IR (film) m = 1740
´
chromatography was performed on Silicycle 60 (Que-
1
bec, Qc, Canada) 230–400 mesh silica gel. Thin-layer
chromatography (TLC) was performed on 0.25 mm
silica gel 60 F254 plates (Whatman, Madison, UK)
and compounds were visualized by exposure to UV
light (254 nm) or a heated solution of ammonium/sul-
furic acid/water. Infrared (IR) spectra were recorded
on a Perkin-Elmer series 1600 FT-IR spectrometer
(Norwalk, CT, USA) and only the significant bands
(C@O, ketone); H NMR (400 MHz, CDCl3) d = 0.83
(s, 18-CH3), 0.85 (s, 19-CH3), 0.70–2.00 (m, 20 H),
2.07 (m, 16a-CH), 2.43 (dd, J1 = 8.8 Hz, J2 = 19.2,
16b-CH), 3.93 (s, OCH2CH2O); 13C NMR (75.5 MHz,
CDCl3) d = 11.5, 14.0, 20.6, 21.9, 28.4, 30.9, 31.2, 31.7,
35.2, 35.8, 36.0, 36.1, 38.1, 43.8, 47.9, 51.6, 54.3, 64.3
(2·), 109.4, 221.5; LRMS for C21H33O3 [MH]+: 333.1
m/z.
reported in cmꢀ1
.
1H and 13C spectra were recorded
with Brucker AC/F300 spectrometer (Billerica,
a
4.1.4. Synthesis of 3-dioxolane-16a-allyl-5a-androstan-
17-one (8). A solution of diisopropylamine (2.35 mL,
16.8 mmol) in dry THF was stirred under argon at
0 ꢁC and a 2.4-M solution of butyllithium in hexanes
(7.0 mL, 16.8 mmol) was added dropwise. After
30 min, ketone 7 (5.1 g, 15.3 mmol) dissolved in dry
THF was added dropwise in the resulting lithium diiso-
propylamine (LDA) solution. This mixture was allowed
to stir for 30 min at 0 ꢁC, then cooled at ꢀ78 ꢁC and al-
lyl bromide (1.45 mL, 16.8 mmol) was added dropwise.
The reaction mixture was stirred for 24 h from ꢀ78 to
0 ꢁC. Water was added to quench the reaction and the
crude product was extracted with EtOAc. The organic
phase was washed with brine, dried over MgSO4, and
evaporated under reduced pressure. Purification by flash
chromatography (hexanes/EtOAc, 92:8) afforded the
major 16a-isomer 8 containing only 1.5% of minor
16b-isomer (2.2 g, 38%) and starting material (60%).
IR (film) m = 1740 (C@O, ketone); 1H NMR
(400 MHz, CDCl3) d = 0.83 (s, 18-CH3), 0.90 (s, 19-
CH3), 0.70–1.85 (m, 20H), 2.05 (m, 1 H of CH2-10),
2.50 (m, 16b-OH and 1 H of CH2-10), 3.93 (s, OCH2-
CH2O), 5.03 (m, CH2-30), 5.77 (m, CH-20); 13C NMR
(75.5 MHz, CDCl3) d = 11.5, 14.7, 20.5, 26.9, 28.4,
30.8, 31.2, 31.7, 35.1, 35.2, 35.8, 36.1, 38.0, 43.7, 44.2,
48.6, 49.1, 54.3, 64.3 (2·), 109.4, 116.5, 136.6, 222.0;
LRMS for C24H37O3 [MH]+: 373.1 m/z.
MA, USA) at 300 and 75.5 MHz, respectively, and a
Bruker AVANCE 400 spectrometer at 400 (1H) and
100.6 (13C) MHz. The chemical shifts (d) are expressed
in ppm and referenced to chloroform (7.26 and
77.16 ppm), acetone (2.05 and 206.26 ppm) or metha-
1
nol (3.31 and 49.0 ppm) for H and 13C, respectively.
Low-resolution mass spectra (LRMS) were recorded
with an LCQ Finnigan apparatus (San Jose, CA,
USA) equipped with an atmospheric pressure chemical
ionization (APCI) source on positive mode. High-res-
olution mass spectra (HRMS) were provided by the
Regional Laboratory for Instrumental Analysis (Uni-
´
´
´
versite de Montreal, Montreal, Qc, Canada). High-
performance liquid chromatography (HPLC) analyses
were carried out using a Waters Associates System
(Waters 600E Pump, Waters 717 plus Autosampler,
Waters 996 Photodiode Array Detector: 200-400 nm)
(Milford, MA, USA), a Nova-Pack C18 column
˚
(150 · 3.9 mm, 4 lm, 60 A), and a solution of MeOH
containing 20 mM AcONH4 as eluent (1 mL/min flow
rate). The analysis was only possible for a compound
having a ketone, an iodide, an azide or a phenyl
group.
4.1.2. Synthesis of 3-dioxolane-5a-androstan-17b-ol
(6).47 To a solution of DHT (5) (5.1 g, 17.7 mmol) in
dry benzene (80 mL) were added under an argon atmo-
sphere ethylene glycol in excess (29.6 mL, 530 mmol)
and p-TSA (33.6 mg, 0.177 mmol). The mixture was re-
fluxed for 24 h using a Dean–Stark/water separator. The
solution was quenched by the addition of saturated
aqueous NaHCO3 and concentrated under vacuum.
The product was extracted with EtOAc, washed with
brine, dried over MgSO4, and evaporated to dryness.
Flash chromatography (hexanes/EtOAc, 95:5) afforded
6 as a white solid (5.2 g, 88%). IR (film) m = 3384 (OH,
4.1.5. Synthesis of 3-dioxolane-16a-allyl-5a-androstan-
17b-ol (9). A solution of ketone 8 (2.2 g, 5.8 mmol) in
dry THF (50 mL) and LiAlH4 (329.4 mg, 8.7 mmol)
was stirred under argon at ꢀ78 ꢁC. After 10 h, the reac-
tion was quenched by the addition of water and 15%
aqueous NaOH solution. Rochelle salt (100 mL, 1 M
solution in water) was added and the mixture extracted
with EtOAc. The combined organic layer was washed
with brine, dried over MgSO4, filtered, and evaporated
under reduced pressure. Purification by flash chroma-
tography (hexanes/EtOAc, 85:15) yielded 2.0 g (92%)
of alcohol 9. IR (film) m = 3358 (OH); 1H NMR
(400 MHz, (CD3)2CO) d = 0.77 (s, 18-CH3), 0.84 (s,
19-CH3), 0.60–1.90 (m, 20H), 1.99 and 2.36 (2 m,
CH2-10), 3.18 (dd, J1 = 5.6 Hz, J2 = 7.6 Hz, CH-17a),
3.60 (d, J = 5.4 Hz, 17b-OH), 3.87 (s, OCH2CH2O),
1
alcohol); H NMR (400 MHz, CDCl3) d = 0.73 (s, 18-
CH3), 0.82 (s, 19-CH3), 0.85–2.15 (m, 23H), 3.63 (t,
J = 8.5 Hz, 17a-CH), 3.94 (s, OCH2CH2O); 13C NMR
(75.5 MHz, CDCl3) d = 11.3, 11.6, 20.9, 22,8, 23.5,
28.6, 30.7, 31.3, 31.6, 35.7, 36.2, 36.9, 38.1, 43.1, 43.9,
51.1, 54.3, 64.1 (2·), 82.0, 109.3; LRMS for C21H35O3
[MH]+: 335.2 m/z.

3012
J. Roy et al. / Bioorg. Med. Chem. 15 (2007) 3003–3018
4.98 (m, CH2-30), 5.85 (m, CH-20); 13C NMR
(75.5 MHz, CDCl3) d = 11.6, 12.1, 20.8, 28.6, 30.1,
31.3, 31.6, 35.4, 35.7, 36.1, 36.9, 38.1, 39.8, 43.4, 43.9,
44.0, 49.4, 54.3, 64.3 (2·), 87.6, 109.5, 115.8, 138.1;
LRMS for C24H39O3 [MH]+: 375.1 m/z.
(10 mL) was slowly added with a syringe and the reac-
tion mixture was stirred for 2 h at rt. An aqueous satu-
rated NaHCO3 solution was then added and the
reaction mixture was extracted with EtOAc. The organic
phase was dried over MgSO4 and solvent evaporated
under reduced pressure. The solution was concentrated
under reduced pressure and extracted with CH2Cl2.
The organic phase was dried over MgSO4 and the crude
product was purified by flash chromatography with hex-
anes/EtOAc (80:20) as eluent to yield diol 12 (688 mg,
70%). IR (film) m = 3396 and 3223 (OH), 1707 (C=O);
1H NMR (400 MHz, CD3OD) d = 0.79 (s, 18-CH3),
1.07 (s, 19-CH3), 0.75–2.60 (m, 25H), 3.13 (d,
4.1.6. Synthesis of 3-dioxolane-16a-(30-hydroxypropyl)-
5a-androstan-17b-ol (10). To a stirred solution of al-
kene 9 (113 mg, 0.30 mmol) in dry THF (15 mL) at
0 ꢁC was added dropwise a 1-M borane solution in
THF (1.81 mL, 1.81 mmol). The mixture was allowed
to react under argon for 5 h, then a 1-N aqueous
NaOAc solution (1.81 mL) and H2O2 (33% w/v,
0.05 mL) were added and the resulting mixture was
stirred at rt for 1 h. The reaction was quenched by
addition of water, extracted with EtOAc, and the or-
ganic phase dried over MgSO4. After evaporation un-
der reduced pressure, the crude product was purified
by flash chromatography using hexanes/EtOAc
(70:30) as eluent to give diol 10 (86 mg, 73%). IR
J = 7.7 Hz,
CH-17a),
3.55
(dd,
J1 = 5.6 Hz,
J2 = 6.6 Hz, CH2-OH); 13C NMR (75.5 MHz, CD3OD)
d = 11.7, 12.6, 22.0, 30.0, 31.4, 32.5, 33.1, 36.6, 36.9,
38.1, 38.9, 39.8, 43.8, 45.1, 45.5, ꢁ49 (under solvent
peaks), 50.4, 55.4, 63.2, 88.8, 214.8; LRMS for
C22H37O3 [MH]+: 349.1 m/z; HRMS: calcd for
C22H36O3 + Na [M+Na]+ 371.25567, found 371.25554;
HPLC purity of 99% (RT = 27.8 min).
1
(film) m = 3256 (OH); H NMR (400 MHz, (CD3)2CO)
d = 0.76 (s, 18-CH3), 0.84 (s, 19-CH3), 0.67–1.79 (m,
25H), 3.12 (dd, J1 = 5.4 Hz, J2 = 7.5 Hz, CH-17a),
3.41 (t, J = 5.2 Hz, CH2-OH), 3.53 (dd, J1 = 1.3 Hz,
J2 = 4.9 Hz, CH2-OH), 3.58 (d, J = 5.3 Hz, 17b-OH),
3.86 (m, OCH2CH2O); 13C NMR (75 MHz,
(CD3)2CO) d = 11.9, 12.7, 21.6, ꢁ30 (under solvent),
31.4, 32.1, 32.5, 32.8, 33.2, 36.3, 37.0, 38.0, 38.9,
44.0, 44.6, 44.9, 50.4, 55.6, 63.0, 64.8 (2·), 88.4,
109.6; LRMS for C24H41O4 [MH]+: 393.0 m/z.
4.1.9. Synthesis of 16a-(30-iodopropyl)-17b-hydroxy-5a-
androstan-3-one (13). Diol 12 (85 mg, 0.24 mmol) was
dissolved in dry CH2Cl2 (10 mL) and PPh3 (95 mg,
0.36 mmol). I2 (92 mg, 0.36 mmol) and imidazole
(43 mg, 0.63 mmol) were added and the solution was
stirred at rt under argon for 2 h. The mixture was
quenched by addition of water, extracted with CH2Cl2,
and the organic phase dried over MgSO4. After evapo-
ration of solvent under reduced pressure, the crude
product was purified by flash chromatography using
hexanes/EtOAc (80:20) as eluent to yield iodide 13
(61 mg, 55%). IR (KBr) m = 3350 (OH), 1709 (C=O);
1H NMR (400 MHz, CDCl3) d = 0.78 (s, 18-CH3),
1.02 (s, 19-CH3), 0.75–2.50 (m, 26H), 3.20 (m, CH2–I
and CH-17a); LRMS for C22H36O2I [MH]+: 459.1 m/z;
4.1.7. Synthesis of 16a-(30-chloropropyl)-17b-hydroxy-
5a-androstan-3-one (11). Diol 10 (79 mg, 0.20 mmol)
was dissolved in acetone (10 mL) and aqueous (10% v/
v) HCl (7 mL) was slowly added with a syringe, and
the reaction mixture was stirred for 2 h at rt. An aque-
ous saturated NaHCO3 solution was added and the
reaction mixture was extracted with CH2Cl2. The organ-
ic phase was dried over MgSO4 and solvent evaporated
under reduced pressure. This crude keto diol, PPh3
(62 mg, 0.40 mmol), and CCl4 (133 mg, 0.40 mmol) in
dry CH2Cl2 (15 mL) were added at 0 ꢁC and stirred at
rt under argon for 2 h. The reaction mixture was
quenched by addition of water, extracted with CH2Cl2,
and the organic phase dried over MgSO4. After evapo-
ration under reduced pressure, the crude product was
purified by flash chromatography using hexanes/acetone
(80:20) as eluent to yield chloride 11 (36 mg, 49% for
HRMS:
calcd
for
C22H35O2I + Na
[M+Na]+
481.15739, found 481.15733; HPLC purity of 94%
(RT = 33.1 min).
4.1.10. Synthesis of 16a-(30-hydroxypropyl)-5a-andro-
stane-3a,17b-diol (14). Ketone 12 (957 mg, 2.75 mmol)
was dissolved in dry THF (92 mL) and the solution
was cooled at ꢀ78 ꢁC. A 1 M solution of K-Selectride
in THF (3.29 mL, 3.29 mmol) was then added dropwise
under argon. After 1 h, the reaction was quenched by
addition of a saturated aqueous NH4Cl solution
(20 mL) and allowed to warm up to rt. The solution
was concentrated under reduced pressure and the mix-
ture was extracted with CH2Cl2. The organic phase
was dried over MgSO4 and the crude product was puri-
fied by flash chromatography (hexanes/acetone, 80:20)
to afford 14 (481 mg, 50%). IR (KBr) m = 3364 (OH);
1H NMR (400 MHz, CD3OD) d = 0.76 (s, 18-CH3),
0.82 (s, 19-CH3), 0.70–1.85 (m, 25H), 3.12 (d,
J = 7.8 Hz, CH-17a), 3.55 (t, J = 6.0 Hz, CH2–OH),
3.95 (narrow m, CH-3b); 13C NMR (75.5 MHz,
CD3OD) d = 11.7, 12.6, 21.4, 29.6, 29.7, 31.4, 32.5,
32.9, 33.1, 33.5, 36.7, 37.3, 38.2, 40.4, 43.8, 45.1, ꢁ49
(under solvent peaks), 50.7, 56.1, 63.2, 67.2, 88.9; LRMS
for C22H39O3 [MH]+: 351.0 m/z; HRMS: calcd for
C22H38O3 + Na [M+Na]+ 373.27132, found 373.27102.
1
two steps). IR (film) m = 3436 (OH), 1708 (C@O); H
NMR (400 MHz, (CD3)2CO) d = 0.79 (s, 18-CH3),
1.07 (s, 19-CH3), 0.70–2.60 (m, 25H), 3.16 (dd,
J1 = 5.8 Hz, J2 = 7.3 Hz, CH-17a), 3.61 (t, J = 6.7 Hz,
CH2–Cl), 3.67 (d, J = 5.5 Hz, OH); 13C NMR
(75.5 MHz, (CD3)2CO) d = 11.7, 12.6, 21.7, 30.9, 31.3,
32.2, 32.5, 33.9, 36.2, 36.6, 37.8, 38.6, 39.4, 43.4, 44.9,
45.2, 46.2, 47.7, 50.2, 55.0, 88.2, 210.2; LRMS for
C22H35O2 [MꢀCl]+: 331.3 m/z; HRMS: calcd for
C22H36O2Cl [MH]+ 367.23983, found 367.23998; HPLC
purity of 95% (RT = 31.7 min).
4.1.8. Synthesis of 16a-(30-hydroxypropyl)-17b-hydroxy-
5a-androstan-3-one (12). Diol 10 (994 mg, 2.5 mmol) was
dissolved in acetone (100 mL). Aqueous (10% v/v) HCl

J. Roy et al. / Bioorg. Med. Chem. 15 (2007) 3003–3018
3013
4.1.11. Synthesis of 16a-(30-chloropropyl)-5a-androstane-
3a,17b-diol (15). To a solution of triol 14 (117 mg,
0.33 mmol) in dry CH2Cl2 (18 mL) were added PPh3
(350 mg, 1.34 mmol) and CCl4 (205 mg, 1.34 mmol).
After 2 h at rt under argon, the reaction mixture was
quenched by addition of water, extracted with CH2Cl2,
and the organic phase dried over MgSO4. After evapo-
ration under reduced pressure, the crude product was
purified by flash chromatography using hexanes/EtOAc
(80:20) as eluent to yield 15 (92 mg, 75%). IR (film)
extracted with CH2Cl2, and the organic phase dried over
MgSO4. After evaporation under reduced pressure, the
crude product was purified by flash chromatography
using hexanes/acetone (80:20) as eluent to yield 18
(862 mg, 77%). IR (film) m = 3426 (OH), 1705 (C@O);
1H NMR (400 MHz, (CD3)2CO) d = 0.79 (s, 18-CH3),
1.07 (s, 19-CH3), 0.70–2.50 (m, 31H), 3.17 (dd,
J1 = 5.9 Hz, J2 = 7.3 Hz, CH-17a), 3.51 (t, J = 6.8 Hz,
CH2-Br), 3.68 (d, J = 5.5 Hz, OH); 13C NMR
(75.5 MHz, (CD3)2CO) d = 11.8, 12.6, 21.8, ꢁ29 (under
solvent peaks), 30.9, 31.3, 32.2, 32.8, 35.3, 36.2, 37.9,
38.7, 39.4, 39.4, 43.3, 45.3, 47.7, 50.2, 55.1, 88.3, 210.2.
1
m = 3356 (OH); H NMR (400 MHz, CDCl3) d = 0.75
(s, 18-CH3), 0.78 (s, 19-CH3), 0.70–2.00 (m, 25H), 3.19
(d, J = 7.1 Hz, CH-17a), 3.55 (d, J = 6.6 Hz, CH2–Cl),
4.04 (narrow m, CH-3b); 13C NMR (75.5 MHz, CDCl3)
d = 11.4, 12.1, 20.4, 28.5, 29.1, 30.5, 31.7 (2·), 32.3, 33.1,
35.4, 36.0, 36.3, 36.9, 39.3, 43.2, 44.1, 45.5, 49.5, 54.6,
66.7, 88.3; LRMS for C22H38O2Cl - H2O [MHꢀH2O]+:
351.1 m/z; HRMS: calcd for C22H37O2Cl + Na
[M+Na]+ 391.23743, found 391.23768.
4.1.15. Synthesis of 16a-(30-azidopropyl)-17b-hydroxy-
5a-androstan-3-one (19). To a solution of bromide 18
(243 mg, 0.59 mmol) in dry CH3CN (20 mL) was added
NaN3 (154 mg, 2.37 mmol) under an argon atmosphere.
The mixture was stirred at 80 ꢁC for 24 h, and the reac-
tion quenched by addition of water and extracted with
EtOAc. The organic layer was dried over MgSO4 and
evaporated under reduced pressure. Purification of the
crude product by flash chromatography (hexanes/
EtOAc, 80:20) yielded 212 mg (95%) of azide 19. IR
(film) m = 3436 (OH), 2094 (N3), 1711 (C@O); 1H
NMR (400 MHz, (CDCl3) d = 0.79 (s, 18-CH3), 1.02
(s, 19-CH3), 0.70–2.50 (m, 26H), 3.20 (d, J = 7.5 Hz,
CH-17a), 3.29 (t, J = 6.6 Hz, CH2–N3); 13C NMR
(75.5 MHz, CDCl3) d = 11.6, 12.1, 21.0, 27.9, 28.9,
30.5, 31.3, 32.9, 35.3, 35.9, 36.8, 38.3, 38.7, 43.4, 44.1,
44.8, 46.9, 49.3, 51.8, 54.1, 88.2, 212.2; LRMS for
C22H36O2N3 [MH]+: 374.2 m/z; HRMS: calcd for
4.1.12. Synthesis of 16a-(30-fluoropropyl)-5a-androstane-
3a,17b-diol (16). Chloride 15 (23 mg, 0.06 mmol) was
dissolved in dry THF (18 mL) and the solution was
cooled at 0 ꢁC and treated with tetrabutylammonium
fluoride (1 M in THF, 150 lL). This mixture was re-
fluxed overnight and then quenched by addition of
water. The product was extracted with EtOAc and dried
over MgSO4. After evaporation under reduced pressure,
the crude product was purified by flash chromatography
with hexanes/acetone (95:5) to give 16 (16 mg, 73%). IR
(film) m = 3358 (OH); 1H NMR (400 MHz, CD3OD)
d = 0.76 (s, 18-CH3), 0.82 (s, 19-CH3), 0.70–1.90 (m,
25H), 3.12 (d, J = 7.6 Hz, CH-17a), 3.96 (narrow m,
CH-3b), 4.36 (t, J = 5.9 Hz, 1H of CH2–F), 4.48 (t,
J = 6.0 Hz, 1H of CH2–F); 13C NMR (75.5 MHz,
CD3OD) d = 11.7, 12.6, 21.4, 29.6, 29.7, 30.4, 31.3,
32.3, 32.4, 32.9, 33.5, 36.7, 37.2, 38.2, 40.4, 43.7, 45.1,
50.7, 56.1, 67.2, 83.9, 88.9; LRMS for C22H38O2FÆ
H2O [MHꢀH2O]+: 335.2 m/z; HRMS: calcd for
C22H37O2F + Na [M+Na]+ 375.26698, found 375.26678.
C22H35O2N3 + Na
[M+Na]+
396.26215,
396.26185; HPLC purity of 95% (RT = 31.6 min).
found
4.1.16. Synthesis of 16a-(30-azidopropyl)-5a-androstane-
3a,17b-diol (20). Using the same protocol as for the syn-
thesis of 14, ketone 19 was reduced into alcohol 20 and
purified by flash chromatography with hexanes/EtOAc/
CH2Cl2 (80:19:1) to give 63 mg (44%) of 20. IR (KBr)
m = 3450 (OH), 2084 (N3); 1H NMR (400 MHz, CDCl3)
d = 0.76 (s, 18-CH3), 0.78 (s, 19-CH3), 0.75–1.80 (m, 25
H), 3.19 (d, J = 7.4 Hz, CH-17a), 3.29 (t, J = 6.6 Hz,
CH2–N3), 4.04 (narrow m, CH-3b); 13C NMR
(75.5 MHz, CDCl3) d = 11.4, 12.1, 20.3, 27.9, 28.5,
29.1, 30.5, 31.7, 32.3, 32.9, 5.4, 36.0, 36.3, 36.9, 39.3,
43.4, 44.1, 49.5, 51.8, 54.6, 66.7, 88.3; LRMS for
C22H38O2N3 [MH]+: 376.0 m/z; HRMS: calcd for
4.1.13. Synthesis of 16a-(30-iodopropyl)-5a-androstane-
3a,17b-diol (17). Using the same protocol as for the syn-
thesis of 13, the primary alcohol of 14 was substituted
and the compound purified by chromatography with
hexanes/EtOAc (92:8) as eluent to yield 17 (61 mg,
30%). IR (film) m = 3354 (OH); 1H NMR (400 MHz,
CDCl3) d = 0.75 (s, 18-CH3), 0.78 (s, 19-CH3), 0.75–
2.00 (m, 25H), 3.20 (m, CH2–I and CH-17a), 4.15 (nar-
row m, CH-3b); 13C NMR (75.5 MHz, CDCl3) d = 7.5,
11.2, 12.0, 20.2, 28.4, 29.0, 30.4, 31.5, 32.1, 32.4, 35.2,
35.8, 36.2, 36.6, 36.7, 39.2, 42.8, 43.9, 49.3, 54.4, 66.5,
88.1; LRMS for C22H37O2I + NH4 [M+NH4]+: 477.9
m/z; HRMS: calcd for C22H37O2I + Na [M+Na]+
483.17304, found 483.17418; HPLC purity of 99%
(RT = 33.0 min).
C22H37O2N3 + Na
398.27757; HPLC purity of 99% (RT = 31.9 min).
[M+Na]+
398.27780,
found
4.1.17. Synthesis of 16a-(30-aminopropyl)-5a-androstane-
3a,17b-diol (21). To a solution of azide 20 (104 mg,
0.28 mmol) in MeOH (16 mL) was added 10% Pd-C
(20 mg) and the mixture was stirred for 3 h under one
atmosphere of H2. After filtration through Celite, the
solvent was removed under reduced pressure. Purifica-
tion by flash chromatography (acetone/MeOH, 90:10)
afforded 87 mg (90%) of amine 21. IR (KBr) m = 3348
(OH and NH2); 1H NMR (400 MHz, CD3OD)
d = 0.76 (s, 18-CH3), 0.82 (s, 19-CH3), 0.70–1.90 (m,
27H), 2.66 (t, J = 6.0 Hz, CH2–NH2), 3.12 (d,
J = 7.6 Hz, CH-17a), 3.96 (narrow m, CH-3b); 13C
NMR (75.5 MHz, CD3OD) d = 11.7, 12.6, 21.4, 29.6,
4.1.14. Synthesis of 16a-(30-bromopropyl)-17b-hydroxy-
5a-androstan-3-one (18). To a solution of diol 12
(947 mg, 2.72 mmol) in dry CH2Cl2 (18 mL) were added
PPh3 (1.42 g, 5.43 mmol) and CBr4 (1.80 g, 5.43 mmol)
at 0 ꢁC. After 2 h of stirring at rt under argon, the reac-
tion mixture was quenched by addition of water,

3014
J. Roy et al. / Bioorg. Med. Chem. 15 (2007) 3003–3018
29.7, 30.9, 31.3, 32.0, 32.9, 33.5, 33.9, 36.7, 37.2, 38.2,
40.4, 42.5, 43.8, 45.0, 50.7, 56.1, 67.2, 88.8; LRMS for
C22H40O2N [MH]+: 350.3 m/z; HRMS: calcd for
C22H40O2N [MH]+ 350.30536, found 350.30518.
4.1.21. Synthesis of 16a-(30-(400-bromobenzyloxy)propyl)-
17b-hydroxy-5a-androstan-3-one (25). AgOTf (57 mg,
0.22 mmol), 4-bromobenzyl alcohol (56 mg, 0.30 mmol),
and
2,6-di-tert-butyl-4-methyl
pyridine
(91 mg,
0.44 mmol) were dissolved in dry CH2Cl2 (30 mL) and
cooled at 0 ꢁC. To this solution was added bromide 24
(78 mg, 0.15 mmol) and the mixture was stirred for
24 h from 0 ꢁC to rt. After filtration and evaporation
of solvent under reduced pressure, the crude product
was purified by flash chromatography with hexanes/
EtOAc (98:2) to afford the arylether derivative in 76%
yield (71 mg). The TBDMS group of this compound
was hydrolyzed with HCl (2% v/v) in MeOH/CH2Cl2
(1:1). After 2 h at rt, water was added; the methanol
was evaporated under reduced pressure, the aqueous
phase extracted with CH2Cl2 and the organic phase
dried over MgSO4. Flash chromatography using hex-
anes/acetone (90:10) yielded 25 (48 mg, 83%). IR
(KBr) m = 3423 (OH), 1700 (C@O); 1H NMR
(400 MHz, (CD3)2CO) d = 0.79 (s, 18-CH3), 1.07 (s,
19-CH3), 0.75–2.55 (m, 25H), 3.15 (dd, J1 = 5.6 Hz,
J2 = 7.5 Hz, CH-17a), 3.48 (t, J = 6.1 Hz CH2O), 3.62
(d, J = 5.4 Hz, OH), 4.46 (s, OCH2Ph), 7.26 (d,
J = 8.5 Hz, 2H of Ph), 7.47 (d, J = 8.4 Hz, 2H of Ph);
13C NMR (75.5 MHz, (CD3)2CO) d = 11.7, 12.6, 21.7,
ꢁ30 (under solvent peaks), 31.2, 32.2, 33.2, 36.2, 36.6,
37.9, 38.6, 39.4, 43.9, 44.8, 45.2, 47.7, 50.2, 55.1, 71.5,
72.4, 88.3, 121.4, 130.3 (2·), 132.1 (2·), 139.7, 210.2;
LRMS for C29H42O3Br [MH]+: 517.0 and 518.9 m/z;
HRMS: calcd for C29H42O3Br + Na [M+Na]+
539.21313, found 539.21295; HPLC purity of 94%
(RT = 34.5 min).
4.1.18. Synthesis of 16a-(30-thiocyanatopropyl)-17b-
hydroxy-5a-androstan-3-one (22). A mixture of bromide
18 (72 mg, 0.18 mmol) and KSCN (34 mg, 0.36 mmol)
in EtOH (15 mL) was refluxed overnight. The reaction
mixture was quenched by addition of water and ex-
tracted with EtOAc. The combined organic layer was
dried over MgSO4 and evaporated under reduced pres-
sure. Purification of the crude product by flash chroma-
tography (hexanes/EtOAc, 70:30) yielded 56 mg (82%)
of 22. IR (KBr) m = 3519 (OH), 2145 (SC„N), 1705
1
(C@O); H NMR (400 MHz, CDCl3) d = 0.79 (s, 18-
CH3), 1.02 (s, 19-CH3), 0.70–2.50 (m, 25H), 2.97 (t,
J = 7.2 Hz, CH2-SCN), 3.21 (d, J = 7.4 Hz, CH-17a);
13C NMR (75.5 MHz, CDCl3) d = 11.6, 12.1, 21.0,
28.9 (2·), 30.5, 31.3, 34.1, 34.2, 35.3, 35.9, 36.7, 38.2,
38.7, 43.0, 44.1, 44.8, 46.9, 49.3, 54.1, 88.1, ꢁ112 (vw),
212.2; LRMS for C23H36O2NS [MH]+: 390.0 m/z;
HRMS: calcd for C23H35O2NS + Na [M+Na]+
412.22807, found 412.22830; HPLC purity of 98%
(RT = 30.0 min).
4.1.19. Synthesis of 16a-(30-thiocyanatopropyl)-5a-andro-
stane-3a,17b-diol (23). Using the same protocol as for
the synthesis of 14, the ketone of 22 (26 mg, 0.066 mmol)
was reduced and purified using hexanes/EtOAc (70:30)
as eluent to yield 23 (16.2 mg, 62%). IR (KBr)
m = 3475 (OH), 2160 (SC„N); 1H NMR (400 MHz,
CD3OD) d = 0.76 (s, 18-CH3), 0.82 (s, 19-CH3), 0.70–
2.00 (m, 25H), 3.04 (t, J = 7.2 Hz, CH2–SCN), 3.14 (d,
J = 7.4 Hz, CH-17a), 3.96 (narrow m, CH-3b); 13C
NMR (75.5 MHz, CD3OD) d = 11.7, 12.6, 21.4, 29.6,
29.7, 30.1, 31.4, 32.9, 33.5, 35.1, 36.7, 37.2, 38.1, 40.4,
43.4, 45.2, ꢁ49 (under solvent peaks), 50.7, 56.1, 67.2,
4.1.22. Synthesis of 16a-(30-(400-bromobenzyloxy)propyl)-
5a-androstane-3a,17b-diol (26). Using the same protocol
as for the synthesis of 14, ketone 25 (45 mg, 0.087 mmol)
was reduced into alcohol 26 in 2 h. The alcohol was
recovered in the organic layer and purified by flash chro-
matography (hexanes/acetone, 95:5) to afford 26 (27 mg,
88.8,
113.7;
LRMS
for
C23H37O2NS + NH4
1
[M+NH4]+: 409.1 m/z; HRMS: calcd for C23H37O2N-
60%). IR (KBr) m = 3384 (OH); H NMR (400 MHz,
S + Na [M+Na]+ 414.24372, found 414.24366.
(CD3)2CO) d = 0.76 (s, 18-CH3), 0.81 (s, 19-CH3),
0.70–1.85 (m, 25H), 3.14 (dd, J1 = 5.7 Hz, J2 = 7.3 Hz,
CH-17a), 3.29 (d, J = 3.1 Hz, OH), 3.48 (t, J = 6.1 Hz,
CH2O), 3.59 (d, J = 5.4 Hz, OH), 3.93 (m, CH-3b),
4.47 (s, OCH2Ph), 7.31 (d, J = 8.4 Hz, 2H of Ph), 7.52
(d, J = 8.4 Hz, 2H of Ph); 13C NMR (75.5 MHz,
(CD3)2CO) d = 11.8, 12.7, 21.2, ꢁ30 (under solvent
peaks), 30.8, 31.3, 32.8, 33.3, 36.4, 37.1, 38.0, 40.0,
44.0, 44.9, 50.5, 55.9, 66.1, 71.6, 72.4, 88.4, 121.5,
130.4 (2·), 132.2 (2·), 139.8; LRMS for C29H44O3Br
[MH+]: 518.9 and 521.0 m/z; HRMS: calcd for
C29H44O3Br [MH]+ 519.24683, found 519.24662; HPLC
purity of 96% (RT = 36.3 min).
4.1.20. Synthesis of 16a-(30-bromopropyl)-17b-(t-butyl-
dimethylsilyloxy)-5a-androstan-3-one (24). To a solution
of 18 (118 mg, 0.29 mmol) in dry CH2Cl2 (30 mL),
cooled at ꢀ78 ꢁC, were added 2,6-lutidine (0.07 mL,
0.63 mmol) and TBDMS-OTf (0.06 mL, 0.26 mmol).
The reaction mixture was stirred for 3 h under argon,
then quenched by addition of water and 10% HCl.
The acidification step was performed to recover the ke-
tone in position 3. The organic phase was dried over
MgSO4 and evaporated under reduced pressure. Purifi-
cation of the crude compound by flash chromatography
(hexanes/acetone, 95:5) afforded 24 (128 mg, 85%). IR
(film) m = 1715 (C@O); 1H NMR (400 MHz, (CD3)2CO)
d = 0.08 and 0.10 (2s, Si(CH3)2), 0.80 (s, 18-CH3), 0.91
(s, SiC(CH3)3), 1.07 (s, 19-CH3), 0.70–2.50 (m, 40H),
3.26 (d, J = 7.3 Hz, CH-17a), 3.51 (t, J = 6.4 Hz, CH2–
Br); 13C NMR (75.5 MHz, (CD3)2CO) d = ꢀ3.7 and
ꢀ3.6, 11.7, 12.8, 18.8, 21.8, 26.5, ꢁ30 (under solvent
peaks), 32.2, 32.8, 34.1, 35.3, 36.2, 36.6, 38.3, 38.6,
39.4, 44.2, 45.2, 47.7, 49.9, 55.0, 88.8, 210.2; LRMS
for C28H49BrO2Si [MꢀBr]+: 445.2 m/z.
4.1.23. Synthesis of 16a-(30-bromopropyl)-17b-(t-butyl-
dimethylsilyloxy)-5a-androstan-3a-ol (27). Using the
same protocol as for the synthesis of 14, ketone 24 (69
mg) was reduced into alcohol 27. The alcohol was recov-
ered in the organic layer and purified by flash chroma-
tography (hexanes/acetone, 95:5) to afford 27 (40 mg,
58%). IR (film) m = 3352 (OH); 1H NMR (400 MHz,
(CD3)2CO) d = 0.07 and 0.10 (2s, Si(CH3)2), 0.77 (s,
18-CH3), 0.80 (s, 19-CH3), 0.91 (s, SiC(CH3)3), 0.70–

J. Roy et al. / Bioorg. Med. Chem. 15 (2007) 3003–3018
3015
2.00 (m, 40H), 3.23 (m, CH-17a), 3.52 (t, J = 6.6 Hz,
CH2–Br), 3.80 (narrow m, CH-3b); 13C NMR
(75.5 MHz, (CD3)2CO) d = ꢀ3.7 and ꢀ3.6 (Si(CH3)2),
8.4, 11.8, 18.8, 21.2, 26.5, ꢁ30 (under solvent peaks),
32.1, 32.6, 32.8, 33.2, 33.5, 34.1, 35.3, 36.4, 37.0, 38.4,
38.6, 40.0, 44.1, 44.9, 46.2, 50.2, 55.7, 66.1, 88.9; LRMS
for C28H51O2Br [MH]+: 526.1 and 528.2 m/z.
clohexylphosphine[1,3-bis(2,4,6-tri-methylphenyl)-4,5-
dihydroimidazol-2-ylidene][benzylidine]ruthenium(IV)
dichloride) (102 mg, 0.12 mmol). This mixture was re-
fluxed overnight and the solvent was evaporated under
reduced pressure. The crude product was purified by
flash chromatography (hexanes/EtOAc, 95:5) to yield
190 mg (47%) of 31. IR (film) m = 3423 (OH), 1713
(C@O, ester); 1H NMR (400 MHz, (CD3)2CO)
d = 0.76 (s, 1800-CH3), 0.82 (s, 1900-CH3), 0.60–2.50 (m,
23H), 3.21 (d, J = 7.7 Hz, CH-1700a), 3.88 (s, OCH2-
CH2O), 4.77 (d, J = 6.2 Hz, CH2-40), 5.76 and 5.95
(2m, CH-20 and CH-30), 7.52 (m, 2H of Ph), 7.64 (m,
1H of Ph), 8.04 (m, 2H of Ph); LRMS for C32H45O5
[MH]+: 508.9 m/z.
4.1.24. Synthesis of 16a-(30-bromo/chloropropyl)-3a-
methoxy-5a-androstan-17b-ol (28) and 16a-(30-iodopro-
pyl)-3a-methoxy-5a-androstan-17b-ol (29). Bromide 27
(40 mg, 0.08 mmol) was dissolved in dry THF (18 mL)
and the solution was cooled at 0 ꢁC. NaH (18 mg,
0.45 mmol) was added and the mixture was stirred for
30 min before the addition of MeI (43 lL, 0.68 mmol)
and 15-crown-5 ether. The reaction mixture was stirred
at rt for 24 h and then quenched by addition of water.
The product was extracted with CH2Cl2 and dried over
MgSO4. After evaporation under reduced pressure, the
crude product was purified by flash chromatography
with hexanes/EtOAc (98:2) giving a mixture of two
methoxy derivatives (38 mg, 90%). The TBDMS group
of these compounds was removed with a solution of
concentrated HCl in MeOH/CH2Cl2 (1:1), after 2 h,
the reaction was quenched by addition of a saturated
aqueous NaHCO3 solution and extracted with CH2Cl2.
The crude compound was purified by flash chromato-
graphy using hexanes/EtOAc (95:5) to give 28 (21
mg, 70%) and 29. Compound 28: IR (film) m = 3436
(OH); 1H NMR (400 MHz, CDCl3) d = 0.75 (s, 18-
CH3), 0.79 (s, 19-CH3), 0.70–2.00 (m, 25H), 3.18 (d,
J = 7.2 Hz, CH-17a), 3.29 (s, OCH3), 3.48 (m, 1H of
CH2–Cl and CH-3b), 3.55 (t, J = 6.6 Hz, 1H of
CH2Cl); 13C NMR (75 MHz, CDCl3) d = 11.6, 12.1,
20.3, 25.2, 28.6, 30.5, 31.6, 31.9, 32.7, 32.9, 34.4,
35.4, 36.1, 36.9, 39.7, 43.1, 43.2, 44.1, 49.5, 54.5,
55.8, 75.7, 88.3; LRMS for C23H39O2Cl–CH3O [M–
OCH3]+: 351.2 m/z. Compound 29: IR (NaCl)
m = 3474 (OH); 1H NMR (400 MHz, (CDCl3)
d = 0.74 (s, 18-CH3), 0.79 (s, 19-CH3), 0.70–2.00 (m,
25H), 3.20 (m, CH-17a and CH2–I), 3.29 (s, OCH3),
3.42 (m, CH-3b); 13C NMR (75.5 MHz, CDCl3)
d = 7.6, 11.6, 12.1, 20.3, 25.2, 28.6, 30.5, 31.6, 32.6,
32.7, 32.9, 35.4, 36.1, 36.8, 36.9, 39.7, 43.0, 44.1,
49.5, 54.5, 55.8, 75.7, 88.2; LRMS for C23H39O2I–
CH3O [M–OCH3]+: 443.0 m/z; HRMS: calcd for
4.1.27. 16a-(60-chlorohexyl)-17b-hydroxy-5a-androstan-
3-one (32). This compound was prepared in three steps
(16 mg, 18%) as reported below for the synthesis of 33
except that there was no benzoic acid ester to be
hydrolyzed. IR (film) m = 3444 (OH), 1712 (C@O);
1H NMR (400 MHz, (CD3)2CO/D2O) d = 0.78 (s, 18-
CH3), 1.06 (s, 19-CH3), 0.70–2.60 (m, 31H), 3.12 (d,
J = 7.6 Hz, CH-17a), 3.59 (t, J = 6.7 Hz CH2-Cl); 13C
NMR (75.5 MHz, (CD3)2CO) d = 11.7, 12.6, 21.8,
27.7, ꢁ30 (under solvent peaks), 31.2, 32.2, 33.5,
36.2, 36.6, 37.9, 38.6, 39.4, 44.1, 44.8, 45.2, 45.9,
47.7, 50.2, 55.1, 88.3, 210.2; LRMS for C25H41O2
[MꢀCl]+: 373.3 m/z.
4.1.28. Synthesis of 16a-(40-chlorobutyl)-17b-hydroxy-5a-
androstan-3-one (33).
A suspension of alkene 31
(190 mg, 0.37 mmol) and 10% Pd–C (28 mg) in EtOAc
was hydrogenated at one atmosphere for 24 h. After fil-
tration through Celite, the solvent was removed under
reduced pressure. Purification by flash chromatography
(hexanes/EtOAc, 90:10) afforded 157 mg (83%) of
alkane derivative. An aqueous 10% NaOH solution
(8 mL) was added to this intermediate (157 mg,
0.31 mmol) dissolved in methanol (15 mL) and the reac-
tion mixture was stirred for 1.5 h at rt. The product was
extracted with CH2Cl2, the organic phase dried over
MgSO4, and the solvent was removed under reduced
pressure to give the crude diol (124 mg). This diol was
dissolved in acetone (15 mL) and 10% aqueous HCl
(8 mL) was slowly added with a syringe. The reaction
mixture was stirred for 1 h at rt. An aqueous saturated
NaHCO3 solution was then added and the reaction mix-
ture was extracted with CH2Cl2. The organic phase was
dried over MgSO4 and the solvent evaporated under re-
duced pressure. This alcohol (70 mg, 0.17 mmol), PPh3
(101 mg, 0.39 mmol) and CCl4, (62 mg, 0.40 mmol) in
dry CH2Cl2 (10 mL) were then stirred at 0 ꢁC under ar-
gon for 1 h. The reaction was quenched by addition of
water, extracted with CH2Cl2, and the organic phase
dried over MgSO4. After evaporation under reduced
pressure, the crude product was purified by flash chro-
matography using hexanes/EtOAc (80:20) as eluent to
yield chloride 33 (34 mg, 24% for four steps). IR (film)
m = 3420 (OH), 1710 (C@O); 1H NMR (400 MHz,
(CD3)2CO/D2O) d = 0.78 (s, 18-CH3), 1.06 (s, 19-CH3),
0.70–2.60 (m, 27H), 3.13 (d, J = 7.6 Hz, CH-17a), 3.59
(t, J = 6.7 Hz, CH2–Cl); 13C NMR (75.5 MHz,
(CD3)2CO) d = 11.7, 12.6, 21.7, 26.4, 30.6, 31.1, 32.2,
C23H39O2I + Na
497.18845.
[M+Na]+
497.18869,
found
4.1.25. 60-(300-Dioxolane-1700b-hydroxy-500a-androstane-
1600a-yl)-hex-20-enol (30). This compound was prepared
as reported below for the synthesis of 31 in 36% yield
(630 mg). 1H NMR (400 MHz, CDCl3) d = 0.75 (s,
1800-CH3), 0.81 (s, 1900-CH3), 0.60–2.50 (m, 33H), 3.20
(d, J = 7.5 Hz, CH-1700a), 3.65 (m, CH2-OH and OH),
3.93 (s, OCH2CH2O), 5.48 (m, CH-20 and CH-30);
LRMS for C27H44O4 [MH+]: 433.2 m/z.
4.1.26. Synthesis of 40-(300-dioxolane-1700b-hydroxy-500
a-androstan-1600a-yl)-but-20-enyl-benzoate (31). To
a
solution of alkene 9 (300 mg, 0.80 mmol) in dry CH2Cl2
(12 mL) were added benzoic acid allyl ester (390 mg,
2.4 mmol) and 2nd-generation Grubbs catalyst (tricy-

3016
J. Roy et al. / Bioorg. Med. Chem. 15 (2007) 3003–3018
33.7, 35.8, 36.1, 36.6, 37.8, 38.5, 39.3, 43.9, 44.7, 45.2,
45.8, 47.6, 50.2, 55.0, 88.2, 210.2; LRMS for
C23H39O2Cl [M]+: 380.5 m/z.
4.1.32. Synthesis of 16a-(20-bromoethyl)-17b-hydroxy-5a-
androstan-3-one (38). To a solution of alcohol 37 (30 mg,
0.07 mmol) in dry CH2Cl2 (2 mL) under argon were
˚
added molecular sieves (4 A) (50 mg) and 4-meth-
4.1.29. 16a-(60-chlorohexyl)-5a-androstane-3a,17b-diol
(34). This compound was prepared in 75% yield
(15 mg) as reported below for 35. IR (film) m = 3358
ylmorpholine-N-oxide (12 mg, 0.10 mmol) and the mix-
ture was stirred for 15 min at rt. Tetrapropylammonium
perruthenate (2.4 mg, 0.07 mmol) was added and the
solution was stirred for 2 h. The resulting mixture was
filtered on a silica gel column, using hexanes/acetone
(85:15) as eluent to give 29 mg (96%) of the correspond-
ing ketone. Subsequently, a solution of K2CO3 (36 mg,
0.26 mmol) in MeOH (1 mL) was added to a solution
of ketone (29 mg, 0.06 mmol) in MeOH (1 mL) and
the mixture was refluxed for 2 h. The MeOH was evap-
orated under reduced pressure, and the residue dissolved
in EtOAc, and washed with a saturated aqueous NH4Cl
solution. The organic layer was dried over MgSO4 and
evaporated under reduced pressure. Purification of the
crude product by flash chromatography (hexanes/
EtOAc, 85:15) yielded 9 mg (35%) of ketone 38. IR
(film) m = 3422 (OH), 1707 (C@O); 1H NMR
(400 MHz, CDCl3) d = 0.80 (s, 18-CH3), 1.02 (s, 19-
CH3), 0.70–2.50 (m, 23H), 3.22 (d, J = 7.1 Hz, CH-
17a), 3.48 (m, CH2-Br); LRMS for C21H34O2Br
[MH]+: 397.0 and 399.0 m/z.
1
(OH); H NMR (400 MHz, (CD3)2CO) d = 0.76 (s, 18-
CH3), 0.81 (s, 19-CH3), 0.70–1.90 (m, 31H), 3.12 (dd,
J1 = 5.6 Hz, J2 = 7.5 Hz, CH-17a), 3.28 (d, J = 3.1 Hz,
OH), 3.55 (m, CH2-Cl and OH), 3.94 (narrow m, CH-
3b); 13C NMR (75.5 MHz, (CD3)2CO) d = 11.8, 12.7,
21.1, 27.7, 28.9, ꢁ30 (under solvent peaks), 31.2, 32.7,
33.2, 33.5, 36.4, 36.7, 37.0, 38.0, 39.9, 44.1, 44.8, 45.9,
50.5, 55.8, 66.0, 88.4; LRMS for C25H43O2 [MꢀCl]+:
375.3 m/z.
4.1.30. Synthesis of 16a-(40-chlorobutyl)-5a-androstane-
3a,17b-diol (35). Using the same protocol as for the syn-
thesis of 14, chloride 33 (26 mg, 0.06 mmol) was dissolved
in dry THF (2 mL) and the solution cooled at ꢀ78 ꢁC. A
1-M solution of K-Selectride in THF (60 lL, 0.06 mmol)
was added dropwise under argon. After 1 h, the reaction
was quenched by the addition of a saturated aqueous
NH4Cl solution (1 mL). The mixture was concentrated
under reduced pressure and extracted with CH2Cl2. The
organic phase was dried over MgSO4 and the crude prod-
uct was purified by flash chromatography with hexanes/
EtOAc/CH2Cl2 (80:19:1) as eluent to afford 35 (17 mg,
72%). IR (film) m = 3355 (OH); 1H NMR (400 MHz,
(CD3)2CO) d = 0.76 (s, 18-CH3), 0.81 (s, 19-CH3),
0.70–1.90 (m, 27H), 3.28 (dd, J1 = 5.6 Hz, J2 = 7.5 Hz,
CH-17a), 3.28 (d, J = 3.1 Hz, OH), 3.60 (m, CH2–Cl
and OH), 3.93 (narrow m, CH-3b); 13C NMR
(75.5 MHz, (CD3)2CO) d = 11.8, 12.6, 21.1, 26.5, ꢁ30
(under solvent peaks), 31.1, 32.7, 33.2, 33.8, 35.8, 36.3,
37.0, 38.0, 39.9, 44.0, 44.8, 45.9, 50.5, 55.8, 66.0, 88.3;
LRMS for C23H39O2 [MꢀCl]+: 347.3 m/z.
4.1.33. 16a-(20-bromoethyl)-5a-androstane-3a,17b-diol
(39). The TBDMS group of ketone 36 (30 mg,
0.06 mmol) was hydrolyzed with concentrated HCl
(3% v/v) in MeOH. After 2 h at rt, water was added,
the methanol was evaporated under reduced pressure,
the aqueous phase extracted with EtOAc, and the organ-
ic phase dried over MgSO4. Flash chromatography
using hexanes/EtOAc (85:15) yielded 39 (19 mg, 82%).
1
IR (film) m = 3374 (OH); H NMR (400 MHz, CDCl3)
d = 0.76 (s, 18-CH3), 0.78 (s, 19-CH3), 0.70–2.20 (m,
23H), 3.22 (d, J = 7.1 Hz, CH-17a), 3.55 (m, CH2–Br),
4.05 (narrow m, CH-3b); LRMS for C21H35O2BrÆ2H2O
[MHꢀ2H2O]+: 363.1 and 365.1 m/z.
4.1.31. Synthesis of 16a-(20-bromoethyl)-17b-acetoxy-5a-
androstan-3a-ol (37). To a solution of bromide 36
(180 mg, 0.35 mmol) in pyridine (5 mL) were added
Ac2O (357 mg, 0.35 mmol) and DMAP (1.2 mg,
0.01 mmol) under an argon atmosphere. After 2 h at
rt, the reaction was quenched by addition of a saturated
aqueous NaHCO3 solution and the mixture extracted
with EtOAc. The organic phase was washed with a sat-
urated aqueous NaHCO3 solution, dried over MgSO4,
and evaporated under reduced pressure yielding
190 mg of crude intermediate. The TBDMS group of
this intermediate (180 mg) was removed with HF/pyri-
dine (276 lL, 6.48 mmol) in THF at rt for 60 h and
quenched by addition of a saturated aqueous NaHCO3
solution. The reaction mixture was extracted with
EtOAc and the organic layer was washed with a satu-
rated aqueous NaHCO3 solution, dried over MgSO4,
and evaporated under reduced pressure. Purification of
the crude product by flash chromatography (hexanes/
EtOAc, 85:15) yielded 71 mg (49% for two steps) of
alcohol 37. 1H NMR (400 MHz, CDCl3) d = 0.76 (s,
18-CH3), 0.78 (s, 19-CH3), 3.33 (s, COCH3), 0.70–2.30
(m, 32H), 3.35 (m, CH2–Br), 4.03 (narrow m, CH-3b),
4.53 (d, J = 7.6 Hz, CH-17a).
4.2. Assessment of proliferative and antiproliferative
activities on Shionogi (AR+) cells
These cell culture assays were performed as reported in
the literature.27 After a 10-day treatment, the cell growth
was assessed by quantifying DNA content using a mod-
ified Fiszer-Szafarz method,48 as previously described,49
and the results were reported as lg of DNA (Figs. 2 and
4) or percentages (Fig. 3 and Table 1). The antiprolifer-
ative assay on androgen-sensitive (AR+) Shionogi mam-
mary carcinoma cells was carried out at indicated
concentrations of the synthesized compounds and the
results are reported as cell DNA content (lg) or the per-
centage (%) of inhibition relative to the proliferation in-
duced by 0.3 nM DHT. The antiproliferative activity
(%) = 100 · [DNA (DHT)
– DNA (DHT + Com-
pound)/DNA (DHT) ꢀ DNA (Control)]. The prolifera-
tion assay was carried out at indicated concentrations
and results are reported as cell DNA content (lg) or
the percentage (%) of stimulation, taking the stimulation
of androgen DHT (0.3 nM) as 100%. The proliferative
activity (%) = 100 · [DNA (Compound) ꢀ DNA (Con-
trol)/DNA (DHT) ꢀ DNA (Control)].

J. Roy et al. / Bioorg. Med. Chem. 15 (2007) 3003–3018
3017
4.3. Assessment of proliferative activity on PC-3 (ARꢀ)
cells
´
acknowledged. Thanks to Herve Do for the synthesis
of compounds 35–38, to Marie-Claude Trottier for
´
NMR analysis, to Patrick Belanger for LRMS analysis,
4.3.1. Cell culture. PC-3 cells were received from ATCC
and were routinely grown in MEM supplemented with
5% (v/v) fetal bovine serum (FBS), 100 IU penicillin/
mL, 50 lg streptomycin sulfate/ mL, and 1% (v/v) non-
essential amino acids. Cells were incubated at 37 ꢁC in
a humidified atmosphere of 5% CO2 and 95% air. They
were subcultured at near-confluence after gentle diges-
tion in a solution of 0.1% trypsin (Wisent Inc.) in
HEPES buffer containing 3 mM ethylenediaminetetra-
acetic acid (pH 7.2). Cells were pelleted by centrifuga-
tion, resuspended in culture medium, and replated.
and to Diane Michaud for biological evaluations. We
appreciated careful reading of the manuscript by Sylvie
´
Methot.
References and notes
1. Jemal, A.; Murray, T.; Ward, E.; Samuels, A.; Tiwari, R.
C.; Ghafoor, A.; Fewer, E. J.; Thun, M. J. CA Cancer J.
Clin. 2005, 55, 10.
2. McConnell, J. D. Urol. Clin. North Am. 1991, 18, 1.
3. Labrie, F.; Belanger, A.; Luu-The, V.; Labrie, C.; Simard,
J.; Cusan, L.; Gomez, J.; Candas, B. Endocr. Rev. 2005,
26, 361.
4.3.2. Measurement of cell proliferation. Cells at passage
28 were plated in 24-well plates at a density of 7000
cells/well and allowed to adhere to the surface of the
plates for 24 h. Thereafter, the medium was replaced
with fresh medium containing 2% (v/v) charcoal-
stripped FBS and the indicated concentrations of com-
pounds diluted from stock solutions prepared in 99%
redistilled ethanol (a 1000-fold concentrated solution).
Control cells received only the ethanolic vehicle (0.1%
EtOH, v/v). Such a concentration of ethanol does not af-
fect cell growth. The indicated concentrations of com-
pounds were added to triplicate dishes, and cells were
grown for 9 days with changes of medium every 2–3
days. Cell proliferation was determined by measurement
of DNA content as previously described.49
4. Labrie, F. Prostate 2000, 43, 215.
5. Labrie, F.; Candas, B.; Dupont, A.; Cusan, L.; Gomez, J.
L.; Suburu, R. E.; Diamond, P.; Levesque, J.; Belanger, A.
Prostate 1999, 38, 83.
6. Teutsh, G.; Goubet, F.; Battmann, T.; Bonfils, A.;
Bouchoux, F.; Cerede, E.; Gofflo, D.; Kelly, M. G.;
Philibert, D. J. Steroid Biochem. Mol. Biol. 1994, 48, 111.
7. Matias, P. M.; Donner, P.; Coelho, R.; Thomaz, M.;
Peixoto, C.; Macedo, S.; Otto, N.; Joschko, S.; Scholz, P.;
Wegg, A.; Ba¨sler, S.; Scha¨fer, M.; Egner, U.; Carrondo,
M. A. J. Biol. Chem 2000, 275, 26164.
8. Denmeade, S. R.; Isaacs, J. T. Nat. Rev. Cancer 2002, 2,
389.
9. Crawford, E. D.; Eisenberger, M. A.; McLeod, D. G.;
Spaulding, J. T.; Benson, R.; Dorr, F. A.; Blumenstein, D.
A.; Davis, M. A.; Goodman, P. J. N. Engl. J. Med. 1989,
321, 419.
4.4. Steroid receptor affinity assay
10. Crawford, E. D.; Allen, J. A. J. Urol. 1992, 147, 417A.
11. Denis, L. Prostate 1994, 5, 17s.
12. Labrie, F.; Cusan, L.; Gomez, J.; Luu-The, V.; Candas,
B.; Belanger, A.; Labrie, C. J. Steroid Biochem. Mol. Biol.
2004, 92, 327.
13. Labrie, F.; Luu-The, V.; Belanger, A.; Lin, S. X.; Simard,
J.; Pelletier, G.; Labrie, C. J. Endocrinol. 2005, 187, 169.
14. Singh, S. M.; Gauthier, S.; Labrie, F. Curr. Med. Chem.
2000, 7, 211.
The binding affinity assay was performed in human
embryonic kidney (HEK-293) cells stably transfected
with human androgen receptor according to the proce-
50
dure described by Labrie et al. except that the results
were expressed as the percentage of binding affinity at
0.1 and 1 lM of tested compound (the binding affinity
of R1881 (1 lM) was set as 100%).
15. Simard, J.; Singh, S. M.; Labrie, F. Urology 1997, 49, 580.
16. Bennett, C. L.; Tosteson, T. D.; Schmitt, B.; Weinberg, P.
D.; Ernstoff, M. S.; Ross, S. D. Prostate Cancer Prostatic
Dis. 1999, 2, 4.
4.5. Molecular modeling
To predict the optimal position of ligand in the steroid
binding site of human androgen receptor (hAR), we
have used a computational strategy similar to that pro-
posed by Zhorov and Lin.51 The receptor was repre-
sented by a double-shell model based on our hAR
crystallographic structures.43 The inner flexible shell
was composed of hAR ligand binding domain amino
`
17. Labrie, F.; Dupont, A.; Cusan, L.; Lacourciere, Y.;
Monfette, G.; Laberge, J. G.; Edmond, J.; Fazekas, A.
T.; Raynaud, J. P.; Husson, J. M. Clin. Invest. Med. 1982,
5, 267.
18. Denis, L.; Carnelro de Moura, J. L.; Bono, A.; Sylverter,
R.; Whelan, R.; Newling, D.; De Pauw, M. Urology 1993,
42, 119.
19. Dijkman, G. A.; Janknegt, R. A.; Dereijke, T. M.;
Debruyne, F. M. J. J. Urol. 1997, 158, 160.
20. Labrie, F. In Cancer: Principles and practice of oncology;
De Vita, V. T., Hellman, S., Rosenberg, S. A., Eds.;
Lippincott-Raven: Philadelphia, 1999; vol. 13, pp 1–9.
21. Cadilla, R.; Turnbull, P. Curr. Top. Med. Chem 2006, 6,
245.
˚
acids having at least one atom within 8 A of the ligand.
Residue internal coordinates of the flexible shell were
allowed to move during minimization steps. The other
amino acids of the model were included in the outer
rigid shell, in which they were not allowed to vary
during energy minimization.
22. Buijsman, R. C.; Hermkens, P. H.; van Rijn, R. D.; Stock,
H. T.; Teerhuis, N. M. Curr. Med. Chem. 2005, 12, 1017.
23. Hamada, H.; Newmann, F.; Junkman, K. Steroid Acta
Endocrinol. 1963, 44, 380.
24. Labrie, F.; Cusan, L.; Dupont, A.; Gomez, J. L.; Simard,
J.; Luu-The, V.; Pelletier, G.; Labrie, C.; Belanger, A. In
Reproductive Endocrinology, Surgery and Technology;
Acknowledgments
We thank the Canadian Institutes of Health Research
(CIHR) for financial support (D.P.) and scholarship
(J.R.). The contribution of Endorecherche is also

3018
J. Roy et al. / Bioorg. Med. Chem. 15 (2007) 3003–3018
Adashi, E. Y., Rock, J. A., Rosenwaks, Z., Eds.; Lippin-
cott-Raven: Philadelphia, 1996; pp 559–583.
25. Brooks, J. R.; Berman, C.; Primka, R. L.; Reynolds, G.
40. Kaighn, M. E.; Shankar Narayan, K.; Ohnuki, Y.;
Lechner, J. F.; Jones, L. W. Invest. Urol. 1979, 17,
16.
F.; Rasmusson, G. H. Steroids 1986, 47, 1.
´
41. Sack, J. S.; Kish, K. F.; Wang, C.; Attar, R. M.; Kiefer, S.
E.; An, Y.; Wu, G. Y.; Scheffler, J. E.; Salvati, M. E.;
Krystek, S. R., Jr.; Weinmann, R.; Einspahr, H. M. Proc.
Natl. Acad. Sci. U.S.A 2001, 98, 4904.
42. Matias, P. M.; Carrondo, M. A.; Donner, P.; Coelho, R.;
Thomaz, M.; Zhao, X.-Y.; Wegg, A.; Crusius, K.; Egner,
U.; Donner, P. J. Med. Chem. 2002, 45, 1439.
26. Tchedam-Ngatcha, B.; Luu-The, V.; Poirier, D. J. Enz.
Inh. Med. Chem 2002, 17, 155.
27. Maltais, R.; Mercier, C.; Labrie, F.; Poirier, D. Mol.
Diversity 2005, 9, 67.
28. Tremblay, M. R.; Auger, S.; Poirier, D. Bioorg. Med.
Chem. 1995, 5, 505.
´
29. Dionne, P.; Tchedam-Ngatcha, B.; Poirier, D. Steroids
´
ˆ ´
43. Pereira de Jesus-Tran, K.; Cote, P.-L.; Cantin, L.;
Blanchet, J.; Labrie, F.; Breton, R. Protein Sci. 2006, 15,
987.
1997, 10, 674.
30. Bun˜uel, E.; Gil, A. M.; Dıaz-de-Villegas, M. D.; Cativiela,
C. Tetrahedron 2001, 57, 6417.
´
44. Brzozowski, A. M.; Pike, A. C. W.; Dauter, Z.; Hubbard,
¨
R. E.; Bonn, T.; Engstro¨m, O.; Ohman, L.; Greene, G. L.;
Gustafsson, J. A.; Carlquist, M. Nature 1997, 389, 753.
45. So¨derholm, A. A.; Lehtovuori, P. T.; Nyro¨nen, T. H.
J. Med. Chem. 2005, 48, 917.
46. Guex, N.; Peitsch, M. C. Electrophoresis 1997, 18, 2714.
47. Mappus, E.; Renaud, M.; Rolland de Ravel, M.; Grenot,
C.; Cuilleron, C. Y. Steroids 1992, 57, 122.
31. Pouchert, C. J.; Behnke, J., 1st ed. In The Aldrich library
of 13C and 1H FT NMR spectra; Aldrich Chemical
Company, Inc.: St. Louis, 1993; vol. 3, p 564.
32. Blunt, J. W.; Stothers, J. B. Org. Mag. Res. 1977, 9, 439.
33. Trnka, T. M.; Grubbs, R. H. Acc. Chem. Res. 2001, 34, 18.
34. Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org. Lett.
1999, 1, 953.
35. Chatterjee, A. K.; Grubbs, R. H. Org. Lett. 1999, 1, 1751.
36. Neri, R.; Florance, K.; Koziol, P.; Van Cleave, S.
Endocrinology 1972, 91, 427.
37. Poyet, P.; Labrie, F. Mol. Cell. Endocrinol. 1985, 42, 283.
38. Morris, J. J.; Hughes, L. R.; Glen, A. T.; Taylor, P. J.
J. Med. Chem. 1991, 34, 447.
48. Fiszer-Szarfarz, B.; Skafarz, D.; Guevara de Murillo, A.
Anal. Biochem. 1981, 110, 165.
´
49. Simard, J.; Dauvois, S.; Haagensen, D. E.; Levesque,
C.; Merand, Y.; Labrie, F. Endocrinology 1990, 126,
´
3223.
50. Labrie, F.; Luu-The, V.; Calvo, E.; Martel, C.; Cloutier,
´
J.; Gauthier, S.; Belleau, P.; Morissette, J.; Levesque,
M.-H.; Labrie, C. J. Endocinol. 2005, 184, 427.
51. Zhorov, B. S.; Lin, S.-X. Proteins 2000, 38, 414.
39. Wurtz, J. M.; Bourguet, W.; Renaud, J. P.; Vivat, V.;
Chambon, P.; Moras, D.; Gronemeyer, H. Nat. Struct.
Biol. 1996, 3, 87.